But… what is the problem, exactly?
The nightlies are a bit broken due to recent changes and performance optimizations. Sure. This happens from time to time. They are nightly builds, the point is to have code deployed to the machines of early adopters, and bugs identified before we ship a release (even a beta). So it seems the nightly builds are doing their job.
Beta 4 has some difficulty with modern hardware. I know, I bought a new laptop recently and I also have a non-working touchpad on my machine. I just plugged an USB mouse until I have time to fix this.
Waddlesplash not focusing on hardware support? That’s a bit of a ridiculous claim as he single-handedly maintains the entire Wifi stack, and recently started diving into PCI as well to help me with my I2C touchpad driver problems.
Me spending too much time on web browsers? Well, if Haiku did not have an up to date web browser, I wouldn’t use it, and then I would see no use in contributing to it. So, it’s that, or nothing at all. Or maybe I can be convinced to work on something else if I’m getting paid for it (enough to replace my current paid job).
Too many strong personalities in the late 2000s? Looking at Repostat: git repository statistics., and in particular the “author of the year” stats, we can see that in the late 2000s there were about 30 persons contributing to Haiku. This reached a peak of 72 in 2012. To me that does not look like driving people away. And last year we had a “skeleton crew” of 48. Admitedly, that’s less than the number of bones in a typical skeleton, but it’s a reasonable size for an opensource project. If we look at blackduc openhub analysis: Project Summary: Factoids - Open Hub
" Over the past twelve months, 55 developers contributed new code to Haiku. This is one of the largest open-source teams in the world, and is in the top 2% of all project teams on Open Hub."
Regarding the better hardware support in Linux: you may have missed the fact that Intel is now contributing their own drivers directly to Linux, and they are not the only ones to do so. On the other hand, we at Haiku are left on our own to figure things out. I sure would appreciate some help from Intel engineers who know a bit more about the hardware.
Regarding the budget: sure, with more money we could hire more people and do more things. The current budget is ridiculously small and entirely funded by donations. We are not like Mozilla who gets million of dollars from Google, for example. I’m not sure what they do with all this money, but it doesn’t seem quite as efficient.
As for the R1 schedule, well, there are a lot of hard issues on the way, and indeed it would be nice to work on more of them. But there’s only so many of us. But at the same time, there is clearly demand for things that are not at all on the R1 roadmap and would selay it by several months or years, things like 3D acceleration. This demand, however, does not come from the development team, but for the users. And for web browsers, I think the situation is the same.
As a developer, I also have my frustrations with Haiku, with lack of hardware support (I mentioned my touchpad, but before that I also had to fix the PS/2 driver for the keyboard to work, the Wifi required help from Waddlesplash who ported and then updated the driver for my machine from OpenBSD, and then there is the Intel graphics where I’m still trying to figure out how to enable two displays, and then the other little things like reworking and finishing the SD card driver, fixing the soundcard driver, and then having a look at the webcam…). Yes, this is all very annoying and unacceptable. However, I can only work maybe 3-4 hours a day on Haiku at most. And some of these topics require me to learn about the hardware, about the driver architecture, and so on. On the other hand, a WebKit rebuild is something I can do on the background while I watch a movie or so. That explains why, from me, WebKit updates are more frequent than deep dives into drivers.
Well, it’s frustrating for everyone, but eventually the two sides of this equation are related: we try to do things the right way. Otherwise, Haiku would be using the Linux kernel and just accept its quirks and strangeness, and try to keep Tracker and Deskbar running on top of that. But no, we decided to take the more complex route of maintaining our own kernel. Ok, let’s admit it, part of that was because that was an interesting and fun challenge (and it attracted skilled developers who actually pulled it off, and did a decent job at it), but part of it is because we want to maintain control over all we do, and deliver a result that we are proud of. It is unfortunate that this only gets us to “so close!” and not to something people (including ourselves) can actually use. Personally my current solution is to have two machines on my desk, one with Haiku and the other with Windows. Maybe I’ll move the remaining Windows things to an Android tablet instead (with the goal of making it less convenient and forcing me to use Haiku more). Some other people run Haiku in a virtual machine. Few of us really use Haiku as their one and only OS, but it’s not because we’re not interested in it. Most developers are also Haiku users, and the lack of hardware support is as frustrating to us as it is to you. I don’t understand where the impression that users and developers are two completely independant and competing groups comes from?